52 ARVID R. MOLANDER, ALCYONACEA. 
armed anthocodiae, the spicules extend down into the shaft of the anthocodiae, but they 
are never in direct connection with the spicules in the branch bark basal spicules. The 
picules in the anthocodia consist of rods with a few scattered thorns, and, in the branch 
bark, of short rollers with girdles (dumb-bells). Similar spicules are to be found in 
the bark of the trunk, although of smaller size. For the measurements of the spicules 
see the table. In the case of those which are most powerfully armed, somewhat more 
spindle-shaped spicules, which are longer and without regular girdles, are, however, 
to be found in the bark of the branches. In other specimens the mode of growth, as 
well as the armature, differs from the normal, inasmuch as the colonies possess a consi- 
derably slenderer structure, while the branches are more scattered and the armature 
in the anthocodiae is reduced to a few spicules at the base of the tentacles. These 
specimens differ considerably from the normal types and should, as KÖKENTHAL 
remarks, undoubtedly be classed as a new species, if there were not so many transitions 
to the normal type. 
KÖUKENTHAL (1907) indicates the generally distinctive characteristics between Gerse- 
mia rubiformis and G. clavata. He does not point out sufficiently clearly the different 
shape of the basal spicules and the different distribution of spicules within anthocodia 
in G. rubiformis and G. clavata. "The limit between the two species uncertain on that 
account and a later writer, BRocH (1912) has made a mistake regarding these species. 
In his description and illustration of these species he has evidently confused the 
spicules from the upper part of the anthocodia and those from the branch bark of G. 
rubiformis with those of G. clavata, both in the synopsis (p. 4) and in the illustrations 
(nos. 2—4, p. 7 and nos. 6—7, p. 9). In his description of both these species, however, 
he gives in all other respects the exact characteristics. The same remark applies also 
to BrocH (1912) (Duc D ÖRLEANS). It appears as if, in the latter case, BRocH has 
taken a more strongly armed G. rubiformis for a G. clavata. The brief accounts in these 
papers do not allow of a completely reliable decision being come to. 
The profuse development of coenenchym in trunk and branches corresponds to a 
well-developed canal system. Broad, endodermal canals run in a longitudinal direction 
between the coelenterons, and connect these indirectly with each other. In the upper 
part of the branches, where a large number of polyps are packed, the walls between the 
coelenterons are very thin and these walls are often transpierced by direct canals. Lower 
down in the branch, where the number of the coelenterons is smaller, the walls between 
them become thicker and stronger, and direct connection between the coelenterons 
ceases. "The polyps are situated almost exclusevely on the swelled points of the branches, 
and new polyps are not added from the sides of the branches. 
